Commit add9bde2 authored by Heiko Carstens's avatar Heiko Carstens Committed by Martin Schwidefsky

s390/irq: enable irq sum accounting for /proc/stat again

For more than two years, since f2c66cd8
"/proc/stat: scalability of irq num per cpu" the output of /proc/stat is
broken.
The first field in the "intr" line should contain the sum of all interrupts,
however since the above mentioned change it is always zero.

The reason for that is that a per cpu irq sum variable had been introduced
which got incremented when calling kstat_incr_irqs_this_cpu(). However
on s390 we directly incremented only the per cpu per irq counter by accessing
the array element via kstat_cpu(smp_processor_id()).irqs[...].
So fix this and use the kstat_incr_irqs_this_cpu() wrapper which increments
both: the per cpu per irq counter and the per cpu irq sum counter.
Signed-off-by: default avatarHeiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: default avatarMartin Schwidefsky <schwidefsky@de.ibm.com>
parent 1427add0
......@@ -222,7 +222,7 @@ void __irq_entry do_extint(struct pt_regs *regs, struct ext_code ext_code,
/* Serve timer interrupts first. */
clock_comparator_work();
}
kstat_cpu(smp_processor_id()).irqs[EXTERNAL_INTERRUPT]++;
kstat_incr_irqs_this_cpu(EXTERNAL_INTERRUPT, NULL);
if (ext_code.code != 0x1004)
__get_cpu_var(s390_idle).nohz_delay = 1;
......
......@@ -611,7 +611,7 @@ void __irq_entry do_IRQ(struct pt_regs *regs)
tpi_info = (struct tpi_info *)&S390_lowcore.subchannel_id;
irb = (struct irb *)&S390_lowcore.irb;
do {
kstat_cpu(smp_processor_id()).irqs[IO_INTERRUPT]++;
kstat_incr_irqs_this_cpu(IO_INTERRUPT, NULL);
if (tpi_info->adapter_IO) {
do_adapter_IO(tpi_info->isc);
continue;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment